Investigating U.S. Food Insecurity Through Data
Progress Report
1. Administrative Information
- Project: Investigating U.S. Food Insecurity Through Data
- Course: DATA-613 (Fall 2025)
- Instructor: Professor Richard Ressler
- Team: Conrad Muhirwe, Sharon Wanyana, Ryann Tompkins, Alex Arevalo
2. Topic and Data
2.1 Project Focus
Our project examines household food insecurity which is the economic inability to afford adequate food across U.S. counties and states. We link geographic patterns to socioeconomic conditions (income, poverty, unemployment, education) and demographic disparities affecting low-income households, minorities, and families with children.
Objectives:
- Quantify food insecurity prevalence and distribution,
- Analyze socioeconomic relationships,
- Visualize insights via interactive Shiny app for policymakers, researchers, and nonprofit practitioners.
2.2 Literature Review
Our analysis builds on four domains:
- National trends—13.5% U.S. food insecurity in 2023 (U.S. Department of Agriculture, Economic Research Service, 2024)
- Geographic disparities—rural/southern counties show highest rates (Feeding America, 2024)
- Socioeconomic determinants—strong links to income, race, family structure (Gundersen & Ziliak, 2015)
- Policy interventions—SNAP reduces food insecurity ~6% (Gundersen et al., 2017).
2.3 Data Source
Due to the change in project focus from access to insecurity, our data source was updated as well per the distinction below
| Dimension | Food Insecurity (Current Focus) | Food Access (Previous Focus) |
|---|---|---|
| Definition | Economic inability to afford food | Geographic proximity to retailers |
| Question | Can households afford to eat? | How far are stores? |
| Data Source | Feeding America (2009-2023 (Feeding America, 2024) | USDA Atlas (2019) |
2.4 Data Status (75% Complete)
- Completed (2009-2019): Downloaded 11 years of Feeding America county-level data; harmonized variable names across years; cleaned core variables (
food_insecurity_rate,child_food_insecurity_rate,cost_per_meal,food_budget_shortfall); standardized FIPS codes. - In Progress (2020-2023): Integrating recent data with COVID-19 methodology adjustments; adding racial/ethnic disparity measures (2019+); implementing two-dataset strategy for longitudinal trends (2009-2023) and equity analysis (2019-2023).
3. Ethical Review
We adhere to ASA guidelines: using only public, de-identified data (Feeding America, Census); maintaining secure GitHub repository with version control; documenting all transformations in R scripts; correcting data source ensures we measure actual food insecurity (economic capability) not just geographic access, critical for equitable analysis; visualizations will use accessible design and neutral language.
4. Application Design
- Overview: National KPIs, definitions, trends;
- Exploration: Interactive maps (
leaflet), temporal charts (plotly), demographic comparisons;
DT package) per peer review suggestion.- Analysis: Correlation matrices, regression modeling, t-tests.
User Controls: Geographic filters (state/county), temporal sliders (2009-2023), variable selection (socioeconomic indicators), demographic focus (rural/urban, race/ethnicity). Dynamic UI switches between state/county analysis.
User Controls Statistical Tools: DATA-413—correlation, linear regression with diagnostics; DATA-613—classification modeling (high/low food insecurity), model comparison beyond linear regression per team expertise in DATA-427/627.
5. Demonstration Plan
6. Risk Assessment
| Risk | Likelihood | Impact | Mitigation |
|---|---|---|---|
| Variable harmonization errors | Medium | High | Documentation (column_name_years.xlsx), peer review |
| COVID methodology changes | High | Medium | Two-dataset strategy with clear documentation |
| App performance issues | Low | Medium | Data pre-aggregation, reactive filtering |
7. Team Roles and Timeline
| Member | Contributions to Date | Next Phase Responsibilities |
|---|---|---|
| Conrad | Repository setup, app layout design | Statistical modeling, validation |
| Alex | Data acquisition & cleaning scripts (2009-2019) | UI/UX design, data integration (2020-2023) |
| Ryann | Literature & ethical review | Exploratory visualization & Interactive maps |
| Sharon | Variable selection & harmonization strategy | Correlation and regression models |
Conclusion
Updated data source (USDA → Feeding America), harmonized 11+ years of food insecurity data (75% complete), addressed all feedback from the project plan submission & peer review, and established a clear path to an interactive, policy-relevant Shiny application. Our two-dataset strategy handles methodological evolution while maximizing temporal coverage. On track for December 9 delivery.